NOTE: This topic is part of the Uptime Information Hub.
Purpose and Evolution of GNA
With the proliferation of solid state drives (SSDs) in data centers across the world, companies are finding unique, exciting ways to take advantage of the high speed and low latency of SSDs. Within the EMC® Isilon® OneFS® operating system, one of the innovative ways Isilon is using SSDs is for Global Namespace Acceleration (GNA). GNA is a OneFS feature that increases performance across your entire cluster by using SSDs to store file metadata for read-only purposes, even in node pools that don't contain dedicated SSDs.
GNA is managed through the SmartPools™ software module of the OneFS web administration interface. SmartPools enables storage tiering and the ability to aggregate different type of drives (such as SSDs and HDDs) into node pools. When GNA is enabled, all SSDs in the cluster are used to accelerate metadata reads across the entire cluster.
Isilon recommends one SSD per node as a best practice, with two SSDs per node preferred. However, customers with a mix of drive types can benefit from the metadata read acceleration with GNA, regardless of how SSDs are placed across the cluster. When possible, GNA stores metadata in the same node pool that contains the associated data. If there are no dedicated SSDs in the node pool, a random selection is made to any node pool that contains SSDs. Therefore, as long as SSDs are available somewhere in the cluster, a node pool can benefit from GNA.
For more information about GNA, see the "Storage Pools" section of the OneFS Web Administration and CLI Administration Guides.
Upgrading to OneFS 7.0 and Later with GNA Enabled
Starting with OneFS 7.0 and later, the SSD-to-total-capacity minimum ratio that OneFS enforces for Global Namespace Acceleration was raised from 0.5 percent to 1.5 percent. The threshold was raised to help prevent GNA from overloading SSDs. If GNA were to overwhelm the SSDs—too few drives servicing too many requests—the cluster could become unstable. Before you upgrade to OneFS 7.0 or later with GNA enabled, the EMC Isilon cluster must meet the following requirements:
- 20 percent or more of the nodes in the cluster must contain at least one SSD.
- SSDs must make up at least 1.5 percent of the total storage capacity on the cluster; 2 percent storage is recommended.
If your cluster does not meet these system requirements before you upgrade to OneFS 7.0 or later, the upgrade process will be halted, with the following message displayed, where <n%> is the actual SSD percentage of total storage on the cluster.
The upgrade would deactivate Global Namespace Acceleration, as SSD storage is at <n%> of total storage, which is below the new minimum of 1.5%. Either disable GNA, or call Isilon Support. Important: Do NOT bypass this check if you are also failing the CheckDiskpoolsMembership check or any other upgrade check.
Note that in earlier versions of OneFS 7.0 prior to 188.8.131.52, the upgrade completed successfully, but GNA was unexpectedly disabled if SSD storage was lower than the enforced minimum.
Verify that your EMC Isilon Cluster Meets the Minimum Requirements for GNA
Before you upgrade to OneFS 7.0, verify that your EMC Isilon cluster meets the minimum requirements for GNA. Run the following command to view the size and capacity of SSDs on the cluster: isi status –q.
If the cluster does not meet the minimum requirements for the upgrade, do one of the following:
- Increase SSD capacity to meet the requirement
- Disable GNA
If you can't increase SSD capacity or disable GNA, and you'd still like to upgrade, contact Isilon Technical Support for guidance. Our technical support team can help you manage your cluster performance, including:
- Managing large file workflows.
- Storing a smaller number of files—for example, millions rather than billions, so that fewer inodes are in use.
- Limiting by configuration the amount and type of data that uses SSDs.
- Configuring OneFS so that Snapshots or SyncIQ metadata are not stored on SSDs.
For more information, see the EMC Isilon knowledge base article 90371, Upgrade to OneFS 184.108.40.206 and later is halted if Global Namespace Acceleration is enabled and SSD storage is less than the enforced minimum of 1.5 percent of total storage.
Best Practices When Using GNA
Can GNA benefit your workflow? How can you optimize your use of GNA? Here are some important considerations:
- Use GNA for cold data workflows. Certain workflows benefit more from the performance gains of GNA. For example, workflows that require heavy indexing of cold data—archive data stored on disks that is left unmodified for extended periods of time—benefit the most from the increased speed of metadata read acceleration. GNA does not provide any additional benefit to customers who already have solely SSD clusters, because all metadata is already stored on SSDs.
- SSDs must account for a minimum of 1.5 percent of the total space on your cluster. To use GNA, 20 percent of the nodes in your cluster must contain SSDs, and SSDs must account for a minimum of 1.5 percent of the total space on your cluster (2 percent is strongly recommended). This helps to ensure that GNA won't overload the SSDs on your cluster. If you don't maintain these requirements, GNA will be disabled and metadata read acceleration will be lost. To enable GNA again, metadata copies would have to be rebuilt, which takes time.
- Consider how new nodes affect the total cluster space. Adding new nodes to your cluster affects the percentage of nodes with SSDs and total available space on SSDs. Keep this in mind whenever you add new nodes, to avoid disabling GNA and deleting the metadata copy. SSDs must account for a minimum of 1.5 percent of total space on your cluster.
- Do not remove the extra metadata mirror. When GNA is enabled, an SSD is set aside as an additional metadata mirror, along with the existing mirrors set by your requested protection, which is determined in the SmartPools settings. A common misunderstanding is that the SSD is an "extra" mirror that can be safely removed without affecting your cluster. In reality, this extra metadata mirror is critical to the functionality of GNA. Moreover, removing it causes OneFS to rebuild the mirror on another drive. The table to the right shows the number of metadata mirrors required per requested protection level when using GNA.
For more information about requested protection, see the "Storage Pools" section of the OneFS Web Administration Guide.